sophosympatheia/Nova-Tempus-70B-v0.1
Hugging Face
TEXT GENERATIONConcurrency Cost:4Model Size:70BQuant:FP8Ctx Length:32kPublished:Jan 12, 2025License:llama3.3Architecture:Transformer0.0K Warm

Nova-Tempus-70B-v0.1 by sophosympatheia is a 70 billion parameter Llama 3-based merged language model with a 32768 token context length. This uncensored model is specifically designed and optimized for high-quality roleplaying and storytelling tasks, excelling in creative narrative generation. It is a complex merge of several Llama 3 models, including EVA-LLaMA-3.33-70B-v0.1, L3.3-70B-Euryale-v2.3, 70B-L3.3-Cirrus-x1, Anubis-70B-v1, Llama-3.3-70B-Instruct, and Llama-3.1-70B.

Loading preview...

Nova-Tempus-70B-v0.1: A Merged Llama 3 Model for Creative Writing

Nova-Tempus-70B-v0.1 is a 70 billion parameter language model developed by sophosympatheia, built upon the Llama 3 architecture. This model is a complex merge of several Llama 3 variants, including EVA-LLaMA-3.33-70B-v0.1, L3.3-70B-Euryale-v2.3, 70B-L3.3-Cirrus-x1, Anubis-70B-v1, and official Llama-3.3-70B-Instruct and Llama-3.1-70B models. It features a substantial context length of 32768 tokens.

Key Capabilities

  • Optimized for Roleplaying and Storytelling: The model was specifically designed and tested for these creative writing applications.
  • Uncensored Output: Provides uncensored responses, allowing for a wider range of narrative possibilities, with users responsible for content generated.
  • Advanced Merging Technique: Utilizes the Model Stock merge method, combining multiple Llama 3 models to achieve unique characteristics.
  • Customizable Prompting: Includes a detailed 2,500-token system prompt template with "Creative Mindfulness Directives" to guide nuanced and immersive narrative generation, including explicit content and formatting guidelines.

Good For

  • Creative Writers: Ideal for generating detailed and immersive roleplay scenarios and complex story narratives.
  • Developers of Interactive Fiction: Suitable for applications requiring dynamic and uncensored character interactions.
  • Experimentation with Llama 3 Merges: Offers a unique blend of Llama 3 models for those interested in exploring merged model performance in creative tasks.
Popular Sampler Settings

Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.

temperature
top_p
top_k
frequency_penalty
presence_penalty
repetition_penalty
min_p